ScanMate drum scanner DIY maintenance, troubleshooting, mods

monkeyfist said:
Converting to 16bit does not really do that much. It does give you the 16bit color space, but it just interpolates the missing values.

Silverfast? You use silverfast for photo editing?

8 bit is 256 shades per channel, for B&W this literally means 256 shades of grey.

12bit that SM5000 does, is 4096 shades per channel. It's a huge improvement over 8bit. You are dumping a huge amount of color data.

Your comparison does not even show a picture? It's a corner of a negative with nothing on it? Do what you please, but it does not make much sense.

When you scan as tiff, CQ uses the color profile that comes with it. It's set in the CQscan program, it8trsml.icc. That might be cause if you get different colors in tiff than in raw.
Click to expand...

Silverfast or Colorperfect needs 16bit file for editing. Silverfast is nice for inverting negatives, also work well with slides. This is true professional software for working with raw scans.

I know the theory about 8 and 16 bits, but when working with ColorQuarted it's a little different story. With every other scanner I work only on 16bit scans.

My comparison shows a piece of the slide, its dark part. You can see clearly more noise and worse colors on the CQ scan.

No, my Tiff 16 bit scans with CQ do not have it8 profile. I never created any profile of it8 in CQ. There is not even an it8trsml.icc file on my disk. I also see after loading the file that there is no embedded profile.

I made comparison of scans with Scanmate 4000. Dry scan negative Fuji c200. I made one 8 bit scan in ColorTrio and then converted to 16bit before editing. Then scan in CQ tiff 16bit and CQ 8bit converted to 16 bit. For comparison, I also made raw scans using the Pakon scanner and edited the same way as scans from SM. Finally, I added a negative scan directly from the Pakon software.

link to folder https://www.dropbox.com/sh/56uicry4jvs4hxt/AAAO1xpKDqWGW5pIFynDabXra?dl=0


In this comparison, it can be seen that CQ Tiff 8 bit is not the same as CT 8bit. On the 8bit tiff bit converted to 16bit you can see much worse image quality and pasteurization. CT 8bit offers much better color balance and real image sharpness. Unfortunately neither CT nor CQ is able to achieve the quality of the color offered by the Pakon scanner. It may be necessary to apply IT8 calibration, but this is another topic ...

Unfortunately, it is a shame that the Scanmate scanner software is not able to pull out the full potential of these machines.

Ps. If anybody needs raws, i can upload to Dropbox.
 
Thanks for showing us the comparison. I’d love to play with the raw scans (SM4000 and Pakon’s).
 
meloV8 said:
I made comparison of scans with Scanmate 4000. Dry scan negative Fuji c200. I made one 8 bit scan in ColorTrio and then converted to 16bit before editing. Then scan in CQ tiff 16bit and CQ 8bit converted to 16 bit. For comparison, I also made raw scans using the Pakon scanner and edited the same way as scans from SM. Finally, I added a negative scan directly from the Pakon software.
Click to expand...

When you scan in 8bit in CQ, do you make sure to choose the 8bit raw and not the 8bit tiff? There's a huge difference between the two with the raw being much sharper and without all the wierd artifacts such as smearing.

I have often complained about the software's inability to yield a raw 16bit scan, and I still hope a solution will present itself at some point. It seems ridiculous that a scanner that was made for 12bit scans is held back by the software. This was meant to be a professional tool!

I do get excellent scans when scanning in 8bit raw though, no doubt about it. But what if I want to work a lot with the file? 8bit files don't allow great changes to the image.

Edit: btw if you scan at 2000dpi or lower you wont get any artifiacts when using the 16bit tiff option. It seems to me that they applied some sort of automatic blur filter for resolutions over 2000dpi, maybe to counter potential registration issues caused by the increase in drum speed and such? This means that I can get perfect 16bit (12bit really) files when I scan 4x5 at 2000dpi, but that resolution is only really adequate for 4x5 and above. So frustrating.
 
If I use CQ, I only scan in Tiff Raw 8Bit mode. Every other option gives a worse picture quality, and with the Scanmate 3000/4000/5000 scanner, the Raw 16bit tiff simply does not work.

CT and CQ in Tiff Raw 8bit mode produce similar images. However, I prefer to use CT, because in CT I have the ability to select gamma when the file is saved.

Actually, I never scan dry, because I think it's a waste of time. Here I made an exception only for comparison scans. I am writing about this because if I already mount a film on a scanner, then I want to scan it in maximum possible quality. So it always scans wet at maximum resolution. 4000dpi is a good resolution for 35mm and 120 film scans. Also large format 4x5.

Ps. I have added a folder with raw files.
 
Never tried CT. How is the inferface?
I find that 4000dpi is, at times, a bit low for 35mm. With 5000 dpi I get smaller and more defined grain and a bit more detail. I think 6000 dpi might be the sweet spot for 35mm, but I have no way of testing this as I only have the SM5000. For medium format 4000dpi is adequat most of time though. Do you scan 4x5 in 4000 dpi? Must be some really big files then. 2000dpi is fine for Fomapan 100 I think, but my technique when shooting 4x5 could be better ;)
 
ColorTrio is very easy and simply software. It is full working version, no trial, no dongle need. I have dongle of CQ, but for older version. With 5.2 I must turn back the clock in my old XP machine to keep CQ in trial without dongle.

If You make really big prits from 35mm, the more dpi is always better, but in real world the 35mm is less than 20 mpix, specially on color negative or medium iso b&w films, so scanning higher than 3200-4000dpi very rarely shows more real detail, but for big print higher resolution gives more natural look on prints.

I don't scan 4x5 in 4000dpi, but my friend did, and working with such a big file is not easy. SSD hard disk and lot of ram memory is necessary :)
 
If someone is interested in the following ScanMate Plugins please leave a comment or PM:

- ScanView ScanMate PC Plug-In For Photoshop and PhotoStyler 1.3.0.
- ScanView MultiMate for PC 1.1.0.
- ScanView MultiMate for Mac 2.0.

Unfortunately, the 'ScanView ScanMate Mac Plug-In For PhotoShop 1.6.' floppy disk isn't working.
 
I hope I'll find some time to scan the ScanMate 5000 manual soon. There is some info about the plugins. I haven't tested them yet.
 
Kamph said:
Never tried CT. How is the inferface?
I find that 4000dpi is, at times, a bit low for 35mm. With 5000 dpi I get smaller and more defined grain and a bit more detail. I think 6000 dpi might be the sweet spot for 35mm, but I have no way of testing this as I only have the SM5000. For medium format 4000dpi is adequat most of time though. Do you scan 4x5 in 4000 dpi? Must be some really big files then. 2000dpi is fine for Fomapan 100 I think, but my technique when shooting 4x5 could be better ;)
Click to expand...

Always scan for the end use size, saying that you need this and this resolution is wrong approach. As you will not benefit anything from scanning at 4000dpi and then downscaling the picture.
 
monkeyfist said:
Always scan for the end use size, saying that you need this and this resolution is wrong approach. As you will not benefit anything from scanning at 4000dpi and then downscaling the picture.
Click to expand...

Scanning at too low a resolution will lead to certain digital artifacts such as grain aliasing. You could choose a wider aperture to circumvent this of course, but then you loose fine detail.
What's the logic behind scanning for the end use size rather than at 5000 dpi besides file size? Terabytes are cheap these days, so storage really shouldn't be a concern. I would rather have a larger file I could down size to fit all my different printing needs.
 
Kamph said:
Scanning at too low a resolution will lead to certain digital artifacts such as grain aliasing. You could choose a wider aperture to circumvent this of course, but then you loose fine detail.
What's the logic behind scanning for the end use size rather than at 5000 dpi besides file size? Terabytes are cheap these days, so storage really shouldn't be a concern. I would rather have a larger file I could down size to fit all my different printing needs.
Click to expand...

Downscaling a file will lead to the same effect as sharpening. Overall it just degrades the image quality, if you compare it to the same scan scanned to the size used. It will also lead to larger grain, as on a small scan you would not even see the grain. But if you scan at so high resolution that you can see the grain, and then downscale it. The downscaling will try to keep the grain in the photo, as it's just a dumb algorithm calculating what is taken out and what is left in the photo.

You will get more natural look if you scan at the use size.

And it also helps people understand the difference of digital scaling vs what happens in enlargers. That you don't gain anything from having a high dpi scanner, if you are not printing.

Also, i think the grain antialiasing problem is something that drum scanners do not have. CCD scanners always scan at the highest resolution in one direction, as they have fixed width sensor, usually the size of the film its intended for. So the images are always subjected to scaling. So the grain antialiasing is actually probably a result of scanning too large and then downscaling.

This is actually one of the biggest advantages in drums scanners, as they don't have a fixed width CCD sensor. But scan one pixel at the time.
 
I don't think grain aliasing is limited to CCD scanners. If a pixel is bigger than a dye cloud it will not render it accurately. If the pixel is bigger than several dye clouds they will get averaged leading to blotches of color.

You can try it yourself in fact. If you scan a 35mm negative at say 3000 dpi and at 5000 dpi, the higher resolution version will look less grainy - given that they are scanned with the same aperture of course.
I don't see any degradation when down sampling. I will try to test further of course, but I can't see why it would make a difference. Unlike upsizing which requires interpolation downsizing will not add "fake data" but average existing data - not unlike what a scan at a lower resolution would do.
 
Kamph said:
I don't think grain aliasing is limited to CCD scanners. If a pixel is bigger than a dye cloud it will not render it accurately. If the pixel is bigger than several dye clouds they will get averaged leading to blotches of color.

You can try it yourself in fact. If you scan a 35mm negative at say 3000 dpi and at 5000 dpi, the higher resolution version will look less grainy - given that they are scanned with the same aperture of course.
I don't see any degradation when down sampling. I will try to test further of course, but I can't see why it would make a difference. Unlike upsizing which requires interpolation downsizing will not add "fake data" but average existing data - not unlike what a scan at a lower resolution would do.
Click to expand...

I have tested this, downscaling (bicubic smooth gradients in photoshop) will lead to sharpening like effect. Actually, i could use sharpening to produce the exact same look on the picture i had scanned at web use size.

You will always get the best results if you scan at the use size. I'm unsure does this work with CCD scanners, as i came to this conclusion already using a drum scanner. And as CCD scanners have fixed size sensors, it probably does not work on them. Meaning CCD scanners will always end up producing less than optimal results.

That said, i only do this on prints. As i'm only concerned about print quality.

This is probably one of the reason why most film scans in the interwebz look like $hit. With huge grain on small postcard size photos.
 
Looks like you're doing something wrong. No matter what scanner you use, you need to scan at the highest (real) resolution of the scanner if you want to achieve the highest scan quality and print quality.
People are learning their whole life, so everything is in front of You...
 
There is merit to scanning at the desired printing resolution because various printers work best and different resolution. However, with today's Photoshop and sizing algorithms, you are probably best to scan in the highest resolution possible then resizing according to final print needed.

I print on Epson large format printer and I have a method that works for me but I always scan at the highest resolution I will ever needed and downsample from there when needed.

Pali
 
You must log in or register to reply here.
Back
Top Bottom